autonomous formation flying (aff) sensor for precision ... · autonomous formation flying (aff)...
TRANSCRIPT
Autonomous Formation Flying (AFF) Sensor
Autonomous Formation Flying (AFF) Sensor for Precision Formation Flying Missions
MiMi Aung11/21/02
Autonomous Formation Flying (AFF) Sensor
AFF SensorPEM
AFFSystem Engineer
AFFInstrumentLead Engineer
MiMi Aung (335)Jeff Srinivasan (335)
George Purcell (335) Jeff Tien (335)
Antennas
Luis Amaro (336)Tom Osborne (386)Max Vozoff (335)
Consultant:Aluizio Prata (336)
MicrowaveTransceivers and Frequency & Timing
M. Ciminera (333)Gerry Walsh (333)Doug Price (333)Conrad Foster (333) Eric Archer (331)Mary Wells (386)C. Flores-Helizon (386)Tony DeKorte (333)Bud Ansley (333)Mark Fiore (333)
Al Kirk (335)
Prototype Basebandprocessor H/W & SW(modified GRACE processor)
HW: Jeff Tien (335)Garth Franklin (335)
SW: Jeff Srinivasasn (335)Yong Chong (335)Tom Meehan (335)Tim Rogstad (335)Charley Dunn (335)
Integration& Test
Eric Archer (331)Jeff Tien (335)Luis Amaro (336)Larry Young (335)Jeff Srinivasan (335)Phil Mayers (335)Hamid Javadi (336)Randy Bartos (341)Tim Munson (335)Jacob Gorelik (335)Cynthia Lee (335)Garth Franklin (335)George Purcell (335)MiMi Aung (335)Ball Aerospace
60 MHz Basebandprocessor H/W
HW:Bryan Bell (335)Jeff Tien (335) Bob Ahten (344) Alberto Ruiz (335) David Robison (335) Keizo Ishikawa (344) Chuck Lehmeyer (335)Don Nguyen (335)
SW: Yong Chong (335)Ted Stecheson (335)Tim Rogstad (335)Tim Munson (335)Jack Morrison (348)
Analysis & Simulations
George Purcell (335)Larry Young (335)Meera Srinivasan
(331)Kevin Quirk (331)Jeff Srinivasan (335)Jeff Tien (335)
Contributors to the AFF Sensor
StarLightFlight System SE teamBall Aerospace
Autonomous Formation Flying (AFF) Sensor
History and Status� AFF Sensor is a novel design innovated and patented by the JPL GPS team (335)
• The AFF Sensor was initially "seeded" in a small exploratory technology task within the DSN Technology Program (now called the IND Technology Program in 9xx).
• Infused through the New Millennium Program (NMP) for the DS-3 mission (Separated Spacecraft Interferometer).
• Moved into the Origins Program where DS-3 -> ST-3 -> StarLight.• U.S. Patent No. 6,072,433, "An autonomous formation flying sensor for precise autonomous
determination and control of the relative position and attitude for a formation of moving objects", June 6, 2000. (Lawrence E. Young, Stephen M. Lichten, Jeffrey Y. Tien, Charles E. Dunn, Bruce J. Haines, Kenneth H. Lau)
� Technology development activities 1999 - 2002 (StarLight project and Code R funding)
� At this time, a Ka-band prototype of the AFF Sensor has been developed and extensively characterized.� Fundamental algorithms have been demonstrated
� AFF Sensor is ready for adoption into future multiple spacecraft precision formation flying missions
� With customization for individual missions.� Being evaluated further under Terrestrial Planet Finder (TPF) pre-project
technology program
Autonomous Formation Flying (AFF) Sensor
AFF Sensor within a FF Mission� The AFF Sensor is a radio-frequency sensor for multiple spacecraft
precision formation flying (FF) missions. It provides:� Estimates of ranges and bearing angles among multiple spacecraft� A wide field of view for initial acquisition and lost-in-space scenarios.
Directly Nearly Notfacing facing facing(cone< 2°) (2°< cone<45°) (cone>45°)
2 2-30 1601 1-600 5400
*Range (cm)*Bearing angles (arc-minute)
StarLight key performance requirements
StarLight: A separated spacecraft optical interferometer mission
Spacecraft separation: Nominal: 30m � 1000m Recovery capability: 1 - 10 km
*1-σ accuracy
Autonomous Formation Flying (AFF) Sensor
- Performance- (2 cm, 1 arcmin) accuracy when the spacecraft are directly facing each other- Wide field of view coverage (~±70° cone)- 3-D relative positioning (range, azimuth angle, elevation angle)
- Autonomous- No real-time ground-based interaction- Self-contained instrument: Transmit, receive and data communication
HW/SW on multiple spacecraft- No aid from Earth-based GPS system
- Real-time- Real-time determination of range and bearing angles for real-time use in the
formation flying control system
Key Features
Autonomous Formation Flying (AFF) Sensor
Examples of FF MissionsStarLight (flight portion cancelled)
Terrestrial Planet Finder (TPF) ~ 2015
Planet Imager (PI) ~20XX
Laser Interferometer Space Antenna (LISA) ~ 2008
Autonomous Formation Flying (AFF) Sensor
Design Description� An RF instrument that is distributed over multiple s/c.� AFF Sensor on each spacecraft transmits and receives GPS-like signals
S(t) = P(t)D(t)cos(2πft+ φ)where P(t) = ranging code
D(t) = Data bits (telemetry)f = carrier frequency (RF, Ka-band for StarLight)
� 1 TX and 3 RX on the front of each s/c (for determination of range and bearing angles)
� Range is derived mainly from ranging code delay between the s/c� Bearing angles are derived mainly from carrier phase observables� Telemetry exchanged on the RF link
� Calibration across the two spacecraft� Enables each s/c to compute formation flying solutions
AFF Sensor
Spacecraft 2Spacecraft 1
Autonomous Formation Flying (AFF) Sensor
Design Description (Cont�d)
� Signal transmission and reception options� Simultaneously� Time-Division Duplexing (TDD)
� Synchronously� Asynchronously
Autonomous Formation Flying (AFF) Sensor
BPF
AMP
LPF
A/D
Phase Advancer
Code Advancer
NCO
Code Generator
Start/Stop
Real Time Clock
Code Advancer
Code Generator
Start/Stop
Real Time Clock Data
Bit
Extract Phase & Delay
Extract Phase & DelayExtract Data BitAverage Phase & Delay
Compute Sync Time
Tracking Processor
BPF
AMP
LPF
A/D
BPF
AMP
LPF
A/D
BPF
AMP
LPF
A/D
BPF
AMP
BPF
AMP
HPF
LPF
Antenna Subsystem
Microwave Transceiver Subsystem
Baseband Processor Subsystem
Local Oscillator + Reference
AFF Local Estimator &Validation
Spac
ecra
ft C
ompu
ter
AMP
AMP
Split
ter 32.64xxx GHz
Split
ter
32.64 GHz
60 MHz60.xxx MHz
High-SpeedProcessor(FPGA) Command & Data
HandlerCalibration/Compensation
Pseudorange, Phase, Timing
Range & Bearing
InertialAttitude
CommandData
TX Code at 30.xxxMchips/sec
1KBPSMatrixSwitch
Phase & Chip Feedback
60 MHzI,Q
Cos, Sin
E
P
L
LPF
HPF
1PPS
ON, -20dB, <-40dB
Receive Channel
Transmit Channel
Frequency and TimingTransmitters
Receivers
Front Front BackBack
Real, Img
Delay
CorrelationSums
Data
4 2Local Oscillator + Reference
60 MHz
AFF Sensor Subsystems
Autonomous Formation Flying (AFF) Sensor
AFF Sensor on Combiner Spacecraft
AFF Sensor on Collector Spacecraft
Challenges in a Distributed S/C Mission
� To achieve required RF performance in the presence of multipath� Effective antenna pattern� Effective isolation between TX and RX antennas
� To maintain insensitivity to thermal, electrical and mechanical instabilities� Continuous self-calibration techniques across multiple spacecraft
� To implement the required frequency scheme at Ka-band � To operate as a single instrument distributed across multiple spacecraft� Initial signal acquisition and calibration of the distributed system� To be accommodated concurrently with other spacecraft subsystems and the
interferometer, while minimizing multipath
Key challenges are:
Autonomous Formation Flying (AFF) Sensor
Implementation Innovations
� Custom Antenna Design� To minimize multipath while keeping a wide field of view
� Ka-band implementation� Two closely spaced Ka-band references derived from a single (~10
MHz) reference source on each spacecraft� Coherence between RF signals with digital clocks
� Digital Signal Processing� Continuous, instantaneous, self-calibration scheme
� Operates across distributed system� Removes clock offsets, instrumental variations
� Carrier-aided smoothing algorithm to improve range estimates� Coherence of generated code with the RF signals
Autonomous Formation Flying (AFF) Sensor
Technology Development
� An end-to-end Ka-band prototype system was developed.� Related spacecraft mockups were fabricated.� Four testbeds were used.
Key technology challenges have been addressed as follows:
Outdoor Antenna Isolation Testbed
Antenna pattern assessment Testbed
358-meter Range Outdoor Radiated
Testbed
Indoor AFF Sensor Testbed
End-to-end AFF Sensor Error BudgetMax. 1-σ uncertainty: 2 cm (range), 1 arc-minute (bearing)
Analysis
358 m
far-end
near-end
Autonomous Formation Flying (AFF) Sensor
Prototype Baseband Processor –modified GRACE basebandprocessor (IPU)
Prototype Ka-band antenna with choke rings
Ka-band Transmitter:Output: 32.64 GHz RF signal at 13 dBm
Ka-band Receiver:Input 32.64 GHz,Output: 60 MHz 1-bit I and Q
samples
Ka-band Local Oscillator:Output: 32.64 GHz generated from 120 MHz input.
Reference oscillator:120 MHz
Prototype Hardware
Autonomous Formation Flying (AFF) Sensor
60 MHz Baseband Processor
A 60 MHz Baseband Processor will be completed in Q1 FY-03.• Will provide more capability, flexibility and re-programmability for further investigation of the AFF Sensor.
Autonomous Formation Flying (AFF) Sensor
AFF Sensor Antenna Pattern Assessment Testbed
Objective:Evaluate degradation of the delay and phase patterns of the transmitting and receiving antennas due to spacecraft multipath sources.
Approach:� Construct mockups of the AFF mounting plate and sunshades.� Measure the gain and phase patterns of the antennas in the mocked-up flight environment.� Compare measured antenna pattern deviations due to structural environment with the allocation within the end-to-end error budget.
AFF Sensor antennas mounted with mock-ups of the mounting plate and sunshade in the JPL 60-foot anechoic chamber
Autonomous Formation Flying (AFF) Sensor
Antenna Pattern Testing (cont.) - Antenna Plate Baseline
Z
X
Yφφφφ
θθθθ
AFF Sensor Antenna Pattern Assessment Testbed (Cont’d)
Autonomous Formation Flying (AFF) Sensor
Antenna Pattern Testing (cont.) – Collector Shade
AFF Sensor Antenna Pattern Assessment Testbed (Cont’d)
Autonomous Formation Flying (AFF) Sensor
Antenna Pattern Testing (cont.) – Combiner Shade
AFF Sensor Antenna Pattern Assessment Testbed (Cont’d)
Autonomous Formation Flying (AFF) Sensor
θθθθ = 0°(face-on) θθθθ = 0°
(face-on)
φφφφ= 0°(upright)
Rotate test article 90° ccw
φφφφ= 90°(shade on right)
Geometry of Test Setup for Antenna Patterns on the Following Pages
Test Fixture
Upper antenna
Transmitter antenna
Lower antenna
AFF Sensor Antenna Pattern Assessment Testbed (Cont’d)
Upper antenna with sunshade , f = 0°
Lower Antenna gain pattern with sunshade,f = 90°
directly-facingregion
nearly-facingregion
Antenna gain pattern with no sunshade
Ant
enna
gai
n (d
B)
Bearing angle (degree)-5 0 5 45-45 -5 0 5 45-45
-5 0 5 45-45
Ant
enna
gai
n (d
B)
Ant
enna
gai
n (d
B)
Conclusion• Antenna pattern is degraded by the sunshade.• Deviations from the nominal pattern (uncalibrated errors) fed into error trees show that the AFF Sensor can still meet the (2 cm, 1 arc-min) requirement
• Degrades slower than requirement relaxation away from boresight.
AFF Sensor Antenna Pattern Assessment Testbed (Cont’d)
Autonomous Formation Flying (AFF) Sensor
With no sunshade With Combiner s/c sunshadeWith Collector s/c sunshade
AFF Sensor Outdoor Antenna Isolation Testbed
Objective:Determine whether isolation between the transmitting and receiving antennas on the same spacecraft is sufficient and stable.
Approach:� Construct mockups of the mounting plate and sun shades.� Measure isolation between the antennas with and without mocked-up flight environment.
Autonomous Formation Flying (AFF) Sensor
With no sunshade
With Collector s/c sunshade
AFF Sensor Outdoor Antenna Isolation Testbed
Conclusion:� Without the sunshade, the measured levels of isolation matched predicted levels.
� Antenna mounting plate did not introduce any unpredictable effects.
� Sunshade degraded isolation levels.� Level of degradation varied with the shape of the sunshade and with changes in location of the sunshade.� Repeatibility is poor due to effects at the small Ka-band wavelengths.� Multipath sources are localized.
� Possible to control isolation levels by placement of absorber at strategic locations.� Consider Time-division duplexing (TDD) scheme on individual mission basis.
Autonomous Formation Flying (AFF) Sensor
AFF Sensor Indoor Testbed (138-116 Laboratory)
AFF Sensor on Spacecraft #1
AFF Sensor on Spacecraft # 2representative space-loss
Approach:� Integrate an indoor testbed representative of the AFF Sensor distributed on two �spacecraft.� � Composed of Ka-band and digital modules on two sides connected by adjustable waveguideattenuators representative of the space loss. � Each half of the sensor is operated from an independent frequency reference.
Objective:Verify fundamental algorithms distributed across multiple spacecraft.
Autonomous Formation Flying (AFF) Sensor
AFF Sensor Indoor Testbed (Cont’d)
� Continuous self-calibration across two halves was verified.
� Phase observable� Range observable
� Carrier-aided smoothingalgorithm was verified.
� Fundamental, distributed Sensor algorithms have been verified.
� Distributed operation� Ka-band scheme
Phase Observable Calibration
Phas
e (c
ycle
s)
Fig. 6a
10
5
0
-5
Time (seconds)0 2000
Fig. 6c
0
0.05
-0.05
Time (seconds)0 2000
Phas
e (c
ycle
s) Scatter matched SNR predicted
Raw phase
Calibrated phase observable
Time (seconds)Fig. 6f
0
0.02
-0.02
20000 4000
Ran
ge (c
hips
)
Fig. 6d Time (seconds)
20000 4000
0
0.05
-0.05
Ran
ge (c
hips
)
Scatter matched SNR predicted
Raw delay
Calibrated range observable
Range Observable Calibration
Ran
ge (c
hip)
Carrier-aided Smoothing Results
Time (seconds)
0
0.02
-0.02
30001000 4000Summed-channels after smoothing
0
0.02
-0.02
20000 4000
Summed-channels before smoothing
Carrier-aided Smoothing Results
Autonomous Formation Flying (AFF) Sensor
AFF Sensor Indoor Testbed
Conclusion:� The following key technologies were demonstrated in the distributed environment:
� Fundamental AFF Sensor scheme� Continuous self-calibration algorithm operating across two independent halves� Carrier-aided smoothing algorithm requiring sustained coherence across each spacecraft� Basic Ka-band scheme supporting the Sensor design� Time-Division Duplexing (TDD) scheme
Autonomous Formation Flying (AFF) Sensor
End-to-End Functionality Field Test across a 1200-foot Outdoor Range
Objective:� Verify end-to-end functionality of the complete AFF Sensor. (Full performance is not expected in the presence of uncontrolled multipath sources in the outdoor environment.)
Approach:� Operate the prototype AFF Sensor distributed over two halves across a 1200-foot outdoor range.� Introduce changes in ranges and bearing angles.� Derive estimates of the range and bearing angle from observables measured during end-to-end operation.
Autonomous Formation Flying (AFF) Sensor
East End
End-to-End Functionality Field Test across a 1200’ Outdoor Range (Cont’d)
Rcv.Ant. Xmit.
Ant.LocalOscillator
XmitterReceiver
BasebandProcessor
358m Range
RangeChange
(Platform)
Autonomous Formation Flying (AFF) Sensor
West End
End-to-End Functionality Field Test across a 358-m Outdoor Range (Cont’d)
Autonomous Formation Flying (AFF) Sensor
End-to-End Functionality Field Test across a 358-m Outdoor Range (Cont’d)
Conclusion:� End-to-end functionality of the AFF Sensor has been verified by successful operation across the 1200-foot range.
� Range, range-change and bearing angle were determined successfully.� Measured ranges matched the GPS-surveyed �truth� ranges (within the accuracy expected in the presence of uncontrolled multipath)� Range-change and bearing angle estimates matched the �truths� in the experiment.
� Full end-to-end performance needs to be determined by operation across a large (>30 m) range with space-like conditions.
Autonomous Formation Flying (AFF) Sensor
Conclusion� A prototype of the AFF Sensor is fully functional.
� Fundamental algorithms have been verified for operation in a distributedspacecraft environment.
� Performance dependence upon the spacecraft architecture is understood.
� Results show that AFF Sensor can meet the StarLight requirements.
� Is ready for adoption into future multiple spacecraft precision formation flyingmission
� Sensor providing coverage from lost-in-space to full performance at face-to-face spacecraft configuration
� Real-time� Autonomous� Applicable in deep space, near-Earth or regions with no access to GPS� Flexible FPGA-based signal processing� Can be augmented with star-trackers, Global Positioning System receivers
(for near-Earth application)� For each mission, optimize on individual basis by design trade-off among:
spacecraft design, Sensor field of view, formation flying system, instrument design.
Autonomous Formation Flying (AFF) Sensor
Conclusion (Cont�d)Further Investigations for Application to TPF
� Extend for five-spacecraft sensor design� Simultaneous multiple links in a dynamic environment� Which spacecraft are sensing signal from which other spacecraft under what circumstances � Antenna configuration
� Requires instantaneous 4πsteradian coverage
� Much tighter non-directly facing requirements� Multipath modeling and mitigation� Self-Jamming (evaluate TDD for five S/C)� Near-Far issue (jamming from other S/C)
� New signal structure to avoid spacecraft rotation maneuver to resolve bearing angle ambiguities
� Integrated inter-spacecraft communications
Autonomous Formation Flying (AFF) Sensor
Back-up
Autonomous Formation Flying (AFF) Sensor
PARAMETER VALUERF carrier frequency 32.64 GHzChip rate of the PRN ranging code 30 Mchips/sSample rate 60 Msamples/sTransmitted power 20 mW (13 dBm)Transmitting and receiving antenna gain, on axis 9.2 dBiPolarization loss (transmitting linear, receiving circular) 3 dBSky background temperature 3 KReceiver noise temperature 2030 KReceiver noise bandwidth 80 MHzSeparation between spacecraft 30-1000 m
Key Parameters
Autonomous Formation Flying (AFF) Sensor
TIME DIVISION DUPLEXING RANGE ERROR BUDGET, ANY CALIBRATION STATE,range 1000 m, 1-second observables
Error Tree 2
RANGE ESTIMATION ERROR11.1 mm
CARRIER-AIDEDSMOOTHING FOR
40 SEC4.22 mm
UNCALIBRATEDMULTIPATH ON
REMOTE SIGNAL25 mm
STATIC UNCALIBRATED
ERRORS10 mm
TOTAL RANGE MEASUREMENT ERROR27.3 mm
VSWR ATWAVEGUIDE ENDS ?? mm
THERMAL STRUCTURAL
0.2 mm
CALIBRATION-SIGNALERROR2.47 mm
(0.00025 chips)
REMOTE-SIGNALDELAY ERROR
26.7 mm(0.00267 chips)
UNCALIBRATED PATH0.1 mm
RSS
RSS
RSS10
THERMAL NOISE2.4 mm
(0.00024 chips)
OSCILLATORPHASE NOISE
0.603 mm
11
12
14 15
3 4 5 6
UNCALIBRATEDVARIATIONS
0.22 mm
7 8
1
2
16
13
GROUNDCALIBRATIONERROR 5 mm
RSSFILTER
DELAY AS FUNCTION
OF SNR0 mm
OTHER BIASERRORS
0 µm
9
17
OSCILLATORPHASE NOISE
11.6 mm
THERMAL NOISE24 mm
(0.0024 chips)
RSS
18 19
CARRIER-AIDEDSMOOTHING FOR
40 SEC0.39 mm
COVARIANCE
SINGLE-SAMPLE SNR = –15 dBCNR = –3 dB
÷ 40 ÷ 40
FROM ANALYSIS
FROM MEASUREMENT
Autonomous Formation Flying (AFF) Sensor
TIME DIVISION DUPLEXING BEARING-ANGLE ERROR BUDGET,GROUND CALIBRATION + IN-ORBIT ROTATION CALIBRATION,
range 1000 m, 1-second observables
Error Tree 7
TOTAL ERROR, UN-DIFFERENCED PHASE
146.6 µm
REMOTE-SIGNALPHASE ERROR
5.82 µm
UNCALIBRATEDREMOTE- SIGNAL
MULTIPATH48.9 µm
PHASE ERROR(1 ms)
184.2 µm(0.126 rad)
OTHER BIASERRORS
0 µm
ERROR ON ESTIMATEDBEARING ANGLES
az.: 1.00 arcmin, el.: 0.88 arcmin
STAR TRACKER ERROR,EACH AXIS
6 "
2 EPOCHS, ROTATION 100°AROUND LOS
THERMAL NOISE165.8 µm
(0.113 rad)
UNCALIBRATEDSTRUCTURALVARIATIONS
70.7 µm
ERROR (1 ms)17.9 µm
CALIBRATION-SIGNALPHASE ERROR
0.566 µm
RSS
OSCILLATORPHASE NOISE
80.3 µm
RSS THERMAL NOISE17.4 µm
(0.0119 rad)
OSCILLATORPHASE NOISE
4.36 µm
RSS
FILTER DELAYAS FUNCTION
OF SNR
10001 1000
1
10 11
THERMALSTRUCTURAL
50 µmUNCALIBRATED
PATH 50 µm
RSS12 13 14
3 4
5 6 7 8
1
2
9
16
17 18
15
Single-Sample SNR –15 dBCNR –3 dB
COVARIANCE
ALLOWANCE FORUNKNOWN AND
UNDERESTIMATEDERRORS: 118.6 µm
19
FROM ANALYSIS
FROM MEASUREMENT